Section 1.5.1 discussed the binary tense theory proposed by Te Winkel (1866) and Verkuyl (2008), according to which the three binary distinctions in (337) are used in mental representations of tense. Languages may differ when it comes to the grammatical means used for expressing the oppositions in (337): this can be done within the verbal system by means of inflection and/or auxiliaries, but may also involve the use of adverbial phrases, aspectual markers, pragmatic information, etc. Verkuyl claims that Dutch expresses all oppositions in (337) in the verbal system: +past is expressed by inflection, +posterior by means of the verb zullen 'will' , and +perfect by means of the auxiliaries hebben 'to have' and zijn 'to be' .

Section 1.5.2 has argued at length that the claim that zullen is a future auxiliary is incorrect: it is an epistemic modal and it is only due to pragmatic considerations that examples with zullen are sometimes interpreted with future time reference. If this is indeed correct, the Dutch verbal system is based on just the binary features ±past and ±perfect, and therefore does not make an eight-way, but only a four-way tense distinction by means of inflection and auxiliaries . This means that the traditional view on the Dutch verbal tense system in Table 9 from Section 1.5.1, sub I, must be replaced by the one in Table 11; the examples with zullen no longer define a separate set of future tenses.

	present	past
imperfect	simple present (o.t.t.) Ik wandel/Ik zal wandelen. I walk/I will walk	simple past (o.v.t.) Ik wandelde/Ik zou wandelen. I walked/I would walk
perfect	present perfect (v.t.t.) Ik heb gewandeld/ Ik zal hebben gewandeld. I have walked/I will have walked	past perfect (v.v.t.) Ik had gewandeld/ Ik zou hebben gewandeld. I had walked/I would have walked

This revised view on the verbal tense system of Dutch implies that utterances in the simple present/past can normally refer to any event time interval in present/past-tense interval i; eventuality k may precede, follow or overlap with n/n', as indicated in Figure 25. Recall that the number of possible worlds is in principle infinite and that we simply select a number of them that suit our purpose.

The representation of the perfect tenses is virtually identical to that in Figure 25; the only difference is that the eventualities are construed as completed autonomous units within the present/past-tense interval. As before, we indicate this in Figure 26 by means of a vertical line at the end of the event time interval k.

Note that we assumed in the figures above that the default value of time interval j (that is, the time interval within which the eventuality denoted by the lexical projection of the main verb must take place) is equal to that of the complete present/past-tense interval i . In the following sections we will show that contextual information (both of a linguistic and a non-linguistic nature) may overrule this default interpretation and that this gives rise to more restricted interpretations.
      Before we start with the Dutch verbal tense system, we want to note that, although Verkuyl (2008) was probably wrong in assuming that binary tense theory was perfectly mirrored by this system, it seems that Dutch is very suitable for studying the interaction of tense, modality and pragmatic information because it can be characterized as a strongly "tense-oriented" language. First, Dutch normally does not mark mood on the verb (the exception being imperative marking), so that it differs from, e.g., German in that it does not have a productive subjunctive marking on the verb; see Section 1.4.3 for more discussion. Second, Dutch normally does not mark syntactic aspect on the verb, so that it differs from, e.g., English in that progressive aspect can simply be expressed by means of the simple present/past. Third, Dutch does not require epistemic modality to be marked, so that it differs from English in that the expression of non-actualized ("future") events need not be marked by the presence of will (or some other modal verb); Dutch zullen 'will' is optional in such cases. Finally, it may be useful to mention that, contrary to what is the case in English, adverbial phrases like gisteren 'yesterday' that refer to temporal intervals preceding speech time can be used in present-perfect constructions; Dutch does not have the property found in English that such adverbials can only be used in past-tense constructions. As a result, Dutch enables us to directly investigate the interaction of past tense, epistemic modality and pragmatics in deriving special meaning effects without the intervention of any of the more idiosyncratic properties concerning mood/modality, aspect and adverbial modification of the type mentioned above.